Let the american decide

[Platapus]

Originally Posted by Osmium Steele

I know I could find an appropriate sample to show that 60% of americans DO believe that the sun orbits the earth.

For example, I'd only need to find 3 people who already believe it, ask 2 more people who don't, and voila.

Manipulate the sample size/content, and you can "prove" almost anything.

Then that would not be a representative sample and your poll would be biased.

The Gallup organization follows established methodology for selecting poll populations.

Instead of a blanket "you can't trust polls", it would be more appropriate to only trust polls where the methodology is published and to use the results in an appropriate manner.

There is nothing wrong with polling as long as people understand how it was designed, how the data was collected and coded, and how to interpret the results.